Partition-of-Unity Gaussian Kolmogorov–Arnold Networks
Amir Noorizadegan · 2026
Read Partition-of-Unity Gaussian Kolmogorov–Arnold Networks by Amir Noorizadegan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-23599.txt.
Description
Noorizadegan introduces PU-GKAN, a Gaussian Kolmogorov-Arnold network whose edge activations are normalised by their local sum, forming a partition of unity. The design reproduces constants exactly at each edge and admits a clear kernel interpretation, with a practical rule for choosing the scale parameter. Experiments show it is less sensitive to that scale, more accurate on most smooth and rough targets, and more stable in training.
Catalog record
| CREATOR | Amir Noorizadegan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51234 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.23599